ECE/CS 250: Computer Architecture. Basics of Logic Design: Boolean Algebra, Logic Gates. Benjamin Lee

Transcription

1 ECE/CS 250: Computer Architecture Basics of Logic Design: Boolean Algebra, Logic Gates Benjamin Lee Slides based on those from Alvin Lebeck, Daniel Sorin, Andrew Hilton, Amir Roth, Gershon Kedem

2 Admin Resource: Pragmatic Logic by William Eccles In Sakai Resources and linked off web page. Skim for what you need, way too much in there This material is covered in MUCH greater depth in ECE/CS 350 please take ECE/CS 350 if you want to learn enough digital design to build your own processor Download Logisim before Friday s recitation 2

3 What We ve Done, Where We re Going Top Down Application Operating System Software Compiler Firmware CPU Memory I/O system Digital Design Circuit Design Interface Between HW and SW Instruction Set Architecture, Memory, I/O Hardware (Almost) Bottom UP to CPU 3

4 Computer = Machine That Manipulates Bits Everything is in binary (bunches of 0s and 1s) Instructions, numbers, memory locations, etc. Computer is a machine that operates on bits Computers physically made of transistors Electrically controlled switches We can use transistors to build logic E.g., if this bit is a 0 and that bit is a 1, then set some other bit to be a 1 E.g., if the first 5 bits of the instruction are then set this other bit to 1 (to tell the adder to subtract instead of add) 4

5 How Many Transistors Are We Talking About? Pentium III Processor Core 9.5 Million Transistors Total: 28 Million Transistors Pentium 4 Total: 42 Million Transistors Core2 Duo (two processor cores) Total: 290 Million Transistors Core2 Duo Extreme (4 processor cores, 8MB cache) Total: 590 Million Transistors Core i7 with 6-cores Total: 2.27 Billion Transistors How do they design such a thing? Carefully! 5

6 Abstraction! Use of abstraction (key to design of any large system) Put a few (2-8) transistors into a logic gate (or, and, xor, ) Combine gates into logical functions (add, select,.) Combine adders, shifters, etc., together into modules Units with well-defined interfaces for large tasks: e.g., decode Combine a dozen of those into a core Stick 4 cores on a chip 6

7 You are here: Use of abstraction (key to design of any large system) Put a few (2-8) transistors into a logic gate Combine gates into logical functions (add, select,.) Combine adders, muxes, etc together into modules Units with well-defined interfaces for large tasks: e.g., decode Combine a dozen of those into a core Stick 4 cores on a chip 7

8 Boolean Algebra First step to logic: Boolean Algebra Formal apporoach for manipulating of True / False (1/0) After all: everything is just 1s and 0s Boolean Functions: Given inputs (variables): A, B, C, P, Q Compute outputs using logic operators, such as: NOT:!A (= ~A = A = A ) AND: A&B (= A B = A*B = AB = A B) = A&&B in C/Java OR: A B (= A+B = A B) = A B in C/Java XOR: A ^ B (= A B) NAND, NOR, XNOR, Etc. 8

9 Truth Tables Can represent as Truth Table: shows outputs for all inputs a NOT(a)

10 Truth Tables Can represent as truth table: shows outputs for all inputs a NOT(a) a b AND(a,b)

11 Truth Tables Can represent as truth table: shows outputs for all inputs a NOT(a) a b AND(a,b) a b OR(a,b)

12 Truth Tables Can represent as truth table: shows outputs for all inputs a NOT(a) a b AND(a,b) a b OR(a,b) a b XOR(a,b) a b XNOR(a,b) a b NOR(a,b)

13 Boolean Gates (More Later) Gates are electronic devices that implement simple Boolean functions (building blocks of hardware) Examples a b AND(a,b) a b OR(a,b) a NOT(a) a b XOR(a,b) a b NAND(a,b) a b NOR(a,b) a b XNOR(a,b) 13

14 Any Inputs, Any Outputs Can have any # of inputs, any # of outputs Can have arbitrary functions: a b c f 1 f

15 Let s Write a Truth Table for a Function Example: (A & B)!C Start with Empty TT Column Per Input Column Per Output A B C Output 15

16 Let s write a Truth Table for a function Example: (A & B)!C Start with Empty TT Column Per Input Column Per Output A B C Output Fill in Inputs Counting in Binary 16

17 Let s write a Truth Table for a function Example: (A & B)!C Start with Empty TT Column Per Input Column Per Output A B C Output Fill in Inputs Counting in Binary 17

18 Let s write a Truth Table for a function Example: (A & B)!C Start with Empty TT Column Per Input Column Per Output Fill in Inputs Row per set of input values Counting in Binary A B C Output

19 Let s write a Truth Table for a function Example: (A & B)!C Start with Empty TT Column Per Input Column Per Output Fill in Inputs Row per set of input values Counting in Binary Compute Output for reach row (0 & 0)!0 = 0 1 = 1 A B C Output

20 Let s write a Truth Table for a function Example: (A & B)!C Start with Empty TT Column Per Input Column Per Output Fill in Inputs Row per set of input values Counting in Binary Compute Output for each row (0 & 0)!1 = 0 0 = 0 A B C Output

21 Let s write a Truth Table for a function Example: (A & B)!C Start with Empty TT Column Per Input Column Per Output Fill in Inputs Row per set of input values Counting in Binary Compute Output for each row (0 & 1)!0 = 0 1 = 1 A B C Output

22 Let s write a Truth Table for a function Example: (A & B)!C Start with Empty TT Column Per Input Column Per Output Fill in Inputs Row per set of input values Counting in Binary Compute Output for each row A B C Output

23 You try one Try one yourself: (!A B) &!C 23

24 You try one Try one yourself: (!A B) &!C Answer: A B C Output

25 Suppose I turn it around Given a Truth Table, find the formula? Hmmm.. A B C Output

26 Suppose I turn it around Given a Truth Table, find the formula? Hmmm Could write down every true case Then OR together: (!A &!B &!C) (!A &!B & C) (!A & B &!C) (A & B &!C) (A & B & C) A B C Output

27 Suppose I turn it around Given a Truth Table, find the formula? Hmmm.. Could write down every true case Then OR together: (!A &!B &!C) (!A &!B & C) (!A & B &!C) (A & B &!C) (A & B &C) A B C Output

28 Suppose I turn it around Given a Truth Table, find the formula? Hmmm.. Could write down every true case Then OR together: (!A &!B &!C) (!A &!B & C) (!A & B &!C) (A & B &!C) (A & B &C) A B C Output

29 Suppose I turn it around This approach: sum of products Works every time Result is right But really ugly (!A &!B &!C) (!A &!B & C) (!A & B &!C) (A & B &!C) (A & B &C) A B C Output

30 Suppose I turn it around This approach: sum of products Works every time Result is right But really ugly (!A &!B &!C) (!A &!B & C) (!A & B &!C) (A & B &!C) (A & B &C) Could just be (A & B) here? A B C Output

31 Suppose I turn it around This approach: sum of products Works every time Result is right But really ugly (!A &!B &!C) (!A &!B & C) (!A & B &!C) (A&B) A B C Output

32 Suppose I turn it around This approach: sum of products Works every time Result is right But really ugly (!A &!B &!C) (!A &!B & C) (!A & B &!C) (A&B) Could just be (!A &!B) here A B C Output

33 Suppose I turn it around This approach: sum of products Works every time Result is right But really ugly (!A &!B) (!A & B &!C) (A&B) A B C Output

34 Suppose I turn it around This approach: sum of products Works every time Result is right But really ugly (!A &!B) (!A & B &!C) (A&B) Looks nicer Can we do better? A B C Output

35 Suppose I turn it around This approach: sum of products Works every time Result is right But really ugly (!A &!B) (!A & B &!C) (A&B) This has a lot in common:!a & (something) A B C Output

36 Suppose I turn it around This approach: sum of products Works every time Result is right But really ugly (!A &!(B & C)) (A & B) A B C Output

37 Just did some of these by intuition.. but Somewhat intuitive approach to simplifying This is math, so there are formal rules Just like regular algebra 37

38 Boolean Function Simplification Boolean expressions can be simplified by using the following rules (bitwise logical): A & A = A A & 0 = 0 A A = A A 0 = A A & 1 = A A 1 = 1 A &!A = 0 A!A = 1!!A = A & and are both commutative and associative & and can be distributed: A & (B C) = (A & B) (A & C) & and can be subsumed: A (A & B) = A 38

39 DeMorgan s Laws Two (less obvious) Laws of Boolean Algebra: Let s push negations inside, flipping & and!(a & B) = (!A) (!B)!(A B) = (!A) & (!B) You should try this at home build truth tables for both the left and right sides and see that they re the same 39

40 Suppose I turn it around One more simplification on early example: (!A &!(B & C)) (A & B) = (!A & (!B!C)) (A & B) A B C Output

41 Another Simplification Example! (!A!(A & (B C))) 41

42 Simplification Example:! (!A!(A & (B C))) DeMorgan s!!a &!! (A & (B C)) 42

43 Simplification Example:! (!A!(A & (B C))) DeMorgan s!!a &!! (A & (B C)) Double Negation Elimination A & (A & (B C)) 43

44 Simplification Example:! (!A!(A & (B C))) DeMorgan s!!a &!! (A & (B C)) Double Negation Elimination A & (A & (B C)) Associativity of & (A & A) & (B C) 44

45 Simplification Example:! (!A!(A & (B C))) DeMorgan s!!a &!! (A & (B C)) Double Negation Elimination A & (A & (B C)) Associativity of & (A & A) & (B C) A & (B C) A & A = A 45

46 You try this: Come up with a formula for this Truth Table Simplify as much as possible A B C Output

47 You try this: Come up with a formula for this Truth Table Simplify as much as possible Sum of Products: (!A &!B &!C) (!A & B &!C) (A &!B & C) (A & B & C) A B C Output

48 You try this: Simplify first two terms: (!A &!B &!C) (!A & B &!C) 48

49 You try this: Simplify: (!A &!B &!C) (!A & B &!C) Regroup (associative/commutative): ((!A &!C) &!B) ((!A &!C) & B) 49

50 You try this: Simplify: (!A &!B &!C) (!A & B &!C) Regroup (associative/commutative): ((!A &!C) &!B) ((!A &!C) & B) Un-distribute (pull out common factor): (!A &!C) & (!B B) 50

51 You try this: Simplify: (!A &!B &!C) (!A & B &!C) Regroup (associative/commutative): ((!A &!C) &!B) ((!A &!C) & B) Un-distribute: (!A &!C) & (!B B) OR identities: (!A &!C) & true = (!A &!C) 51

52 You try this: Come up with a formula for this Truth Table Simplify as much as possible Sum of Products: (!A &!C) (A &!B & C) (A & B & C) A B C Output

53 You try this: Come up with a formula for this Truth Table Simplify as much as possible Sum of Products: (!A &!C) (A & C) A B C Output

54 Applying the Theory Lots of good theory Can reason about complex Boolean expressions Can design software to minimize But why is this useful? 54

55 Boolean Gates Gates are electronic devices that implement simple Boolean functions (building blocks of hardware) Examples a b AND(a,b) a b OR(a,b) a NOT(a) a b XOR(a,b) a b NAND(a,b) a b NOR(a,b) a b XNOR(a,b) 55

56 Guide to Remembering your Gates This one looks like it just points its input where to go It just produces its input as its output Called a buffer a a 56

57 Guide to Remembering your Gates This one looks like it just points its input where to go It just produces its input as its output Called a buffer a a A circle always means negate a NOT(a) Circle = NOT 57

58 Brief Interlude: Building An Inverter a NOT(a) Vdd = power = 1 P-type: switch is on if input is 0 a NOT(a) N-type: switch is on if input is 1 ground= 0 58

59 Guide to Remembering Your Gates AND Gates have a straight edge, like an A (in AND) Straight like an A a b AND(a,b) OR Gates have a curved edge, like an O (in OR) a b OR(a,b) Curved, like an O 59

60 Guide to Remembering Your Gates If we stick a circle on them a b AND(a,b) a b NAND(a,b) Circle = NOT a b OR(a,b) a b NOR(a,b) We get NAND (NOT-AND) and NOR (NOT-OR) NAND(a,b) = NOT(AND(a,b)) 60

61 Guide to Remembering Your Gates XOR looks like OR (curved line) But has two lines (like an X does) a b XOR(a,b) Can put a dot for XNOR XNOR is 1-bit equals by the way a b XNOR(a,b) 61

62 Boolean Functions, Gates and Circuits Circuits are made from a network of gates. (!A &!C) (A & C) A C Out 62

63 A few more words about gates Gates have inputs and outputs If you try to hook up two outputs, bad things happen (your processor catches fire) a b c d BAD! If you don t hook up an input, it behaves kind of randomly (also not good, but not set-your-chip-on-fire bad) 63

64 Let s Make a Useful Circuit Pick between 2 inputs (called 2-to-1 MUX) Short for multiplexor What might we do first? 64

65 Let s Make a Useful Circuit Pick between 2 inputs (called 2-to-1 MUX) Short for multiplexor What might we do first? Make a truth table? S is selector: S=0, pick A S=1, pick B A B S Output

66 Let s Make a Useful Circuit Pick between 2 inputs (called 2-to-1 MUX) Short for multiplexor What might we do first? Make a truth table? S is selector: S=0, pick A S=1, pick B Next: sum-of-products (!A & B & S) (A &!B &!S) (A & B &!S ) (A & B & S) A B S Output

67 Let s Make a Useful Circuit Pick between 2 inputs (called 2-to-1 MUX) Short for multiplexor What might we do first? Make a truth table? S is selector: S=0, pick A S=1, pick B Next: sum-of-products Simplify (A &!S) (B & S) A B S Output

68 Circuit Example: 2x1 MUX Draw it in gates: A AND MUX(A, B, S) = (A &!S) (B & S) OR output B AND S So common, we give it its own symbol: a b output s 68

69 Example 4x1 MUX a b a 0 c d out b c d y 2 s 0 s 1 S The / 2 on the wire means 2 bits 69

70 Arithmetic and Logical Operations in ISA What operations are there? How do we implement them? Consider a 1-bit Adder 70

71 Summary Boolean Algebra & functions Logic gates (AND, OR, NOT, etc) Multiplexors Download Logisim for Recitation 71